On November 11, 2022, the 43th Zhi·Xin forum of our college was held in the form of online Zoom meeting. Professor Frank Allgöwer from the University of Stuttgart was invited to give a talk titled “Data-driven Model Predictive Control: A Powerful New Approach for a Successful Control Method”.
First, Prof. Frank Allgöwer introduced the design method of feedback controllers in automatic control with an example of a bicycle. Then, model predictive control techniques were discussed by comparing analytical/numerical-based control methods, model predictive control problem formation and solution, and the basic principles of model predictive control were presented. After that, the difficulties of model-based control and the motivation for studying data-driven control were analyzed. In addition, Prof. Frank Allgöwer presented the data-driven model predictive control methods for linear systems in the absence and presence of noisy data in detail. Furthermore, he also introduced the practical applications of data-driven model predictive control methods for nonlinear systems, such as four-tank systems and soft robot systems. Finally, he analyzed many open research problems for data-driven model predictive control techniques in the future.
After the talk,Prof. Frank Allgöwer exchanged views with teachers and students on issues related to data-driven model predictive control. The talk further expanded the research horizons of our students and enhanced their understanding and knowledge of data-driven model predictive control techniques.
Written by Zhang Hao
Photographed by Wang Yunjiao